Volume 11 (2015)
Volume 10 (2014)
Volume 9 (2013)
Volume 5 (2009)
Volume 1 (2005)
Most Downloaded Articles
- Building an NCAA men’s basketball predictive model and quantifying its success by Lopez, Michael J. and Matthews, Gregory J.
- Creating space to shoot: quantifying spatial relative field goal efficiency in basketball by Shortridge, Ashton/ Goldsberry, Kirk and Adams, Matthew
- Predicting the draft and career success of tight ends in the National Football League by Mulholland, Jason and Jensen, Shane T.
- A generative model for predicting outcomes in college basketball by Ruiz, Francisco J. R. and Perez-Cruz, Fernando
- A new approach to bracket prediction in the NCAA Men’s Basketball Tournament based on a dual-proportion likelihood by Gupta, Ajay Andrew
Monte Carlo Simulation for High School Football Playoff Seed Projection
1The College of Wooster
2The College of Wooster
Citation Information: Journal of Quantitative Analysis in Sports. Volume 7, Issue 2, ISSN (Online) 1559-0410, DOI: 10.2202/1559-0410.1330, May 2011
- Published Online:
In Ohio high school football, playoff teams are selected and seeded using an objective point system. Roughly one-fourth of the states teams earn playoff berths, and higher seeds host first-round games. Even in the final week of the season, a teams playoff chances can depend on the outcomes of dozens of other games, making direct computation of playoff probabilities impractical. To make playoff-related predictions, we first estimate win probabilities for all remaining regular-season games by applying a predictive ranking algorithm, then repeatedly simulate the remainder of the regular season. Using the aggregate results, we predict the playoff qualifiers and seeds, and also estimate conditional probabilities (based on the number of future wins) that particular teams earn a berth or a home game. In tracking the results of this model over two seasons, we find that modeling future games substantially increases the accuracy of seed predictions, but adds far less value in predicting the qualifying teams. This phenomenon may be related to the specificity of seed prediction, as compared to the more general nature of predicting a group of teams likely to qualify. That is, the additional information is most useful when making more specific predictions.